MASTER COPY • · go §01UTJHIWIE§1f lfORIE§'f & RANGIE _ _ __ Berkeley , Temperatures in a Large Naturai-Fu CHARLES W.PHI LPOT ABSTRACT: Tempe ratures in a large na tural fuel test fire were meas ure d with bare, shielded aspirated, and shielded unaspirated chromel-alume l t h e r m ocouples. With the bare the rmocouples, values of 2650°F. were r ecorded --much highe r than most previously publi she d data from field and l aboratory wood fires. Soil temperatures were consi st ent with prior work. o o o o o The U. S. Forest Servi c e, u nder contract with the Office of Civil Defens e (Proj ect 2536A), is i nve stigating the characteristics of large mass fires (Countryman 1964) . Knowledge of t emperatures in these fir es is of particular import an c e because it c an help : Establish the r elationship between laboratory and field fires. Test the validity of t h eo retical fire models. Construct thermodynamic models of field s c al e fi res. Compute c onve ction col umn velo citie s. Determine the effect of environment on fire and of fir e on its environment. For su c h purpos es, we need to measur e th e temperatures at several locations within the fire area, the time -temperature relationships, the rate of temp er atur e rise a nd fa 11, t he effe ct of air entrainment on temperatures, and t h e rate of heat loss i nto the soil. Mu ch of the work thus far has been concerned with devising dependable measurement systerns for large - scale t e st fires. This not e describ es methods used and presents some preliminary r e sults. Past Work Flame, gas, and soil temp eratures report e d for fires in natural fuel hav e come mainly from two sources: (a) studies of the effects of fire on the environment and (b) l a b orato r y ex peri men t s concerned with des c ribing the thermodynamic c h aracteristics of fire. Few data seem to be available on ga s temperatur e s from wood fir es . In any case, it is difficult to compare t h e results of most of these experiments owing to differences in quantity and type of fuel s, weather conditions, and lo cation of sample points in and around t he fire. OCD REVIEW NOTICE This report has been r eviewed in th e Office of Ci v il Defens e and approved for publication. Approval does no t s i gn if y that the contents necessa rily reflect the vi ews and polici es o f the Offi ce of Ci v il Defens e . Forest Service - U. S. Department of A g riculture Broido and McMasters (1960), working with a fire that burned 900 tons of windrowed trees and brush, measured temperatures of gases drawn into a pipe through a shielded opening 3 feet above ground. Maximum temperatures recorded inside the pipe were 1720° F. at 1 foot above ground level and 1450° F. at 2 feet above ground. Gas at 1 and 2 feet above ground level inside a similar 3 -foot pipe placed 5 feet out from the fuel reached only 560° F. and 590° F., respectively. However, for a fire in about 250 tons of lumber and railroad ties distributed over 4 acres they reported that gas temperatures inside the pipe reached about 2600° F. before thermocouple failure. Temperatures of 1600° F. were recorded from unshielded thermocouples on 5 headfires and 5 backfires in 8 -year -old gallberry-palmetto roughs on the Alapaha Experimental Range near Tifton, Georgia (Davis and Martin 1960). Nelson and Sims (1934) reported that headfires in the longleaf pine type had temperatures that reached 1500° F. They also judged that by the appearance of clay cones (seger cones), the maximum temperature generated 6 inches above ground in hardwood leaf litter was 1830° F. In the same study, melted or distorted alloy spirals indicated that temperatures at 5 feet above ground generally reached 3 90° F. In a study of bark temperatures during forest fires, Fahnestock and Hare (1961) recorded values from bare iron- constantan thermocouples up to 1555° F., 3 feet above the ground on the lee side of a tree. Vehrencamp (1956) burned railroad ties and reported that temperatures of bare chromel-alumel thermocouples in the center of the convection column at 40 feet reached 1050° F. He also found that a test fire in sagebrush had column temperatures up to 800° F., 20 feet above the fuel bed. His data were collected with bare thermocouples, but Vehrencamp's first approximation of the maximum error due to radiation is 150° F. ,.,J Bruce, Pong, and Fons (1959) measur~q temperatures with bare thermocouples several heights above wood crib fires. At 34 inches above the fuel surface the temperature reached 1500° F.; at 107 inches it rose only to 330° F. The temperatures nearer the fuel were lower than at 34 inches. Beadle (1940) measured soil temperature during forest fires just under the soil surface and l-inch deep. Values of 390° F. and 140° F., respectively, were measured with organic compounds of known melting points. Bentley and Fenner (1958) reported that a fast-moving fire in light litter type fuel heated the soil ''surface" to 200-250° F. and just under the surface to 150-200° F. as measured with temperature:-sensitive paints. Methods The study area is at 7, 100 feet elevation, near Basalt, Nevada. Several plots, ranging in size from 4-1/2 to 50 acres, have been constructed with individual piles of fuel arranged to simulate average American housing tracts (fig. 1). Each pile contains 20 tons (ovendry weight) of natural fuel. About 95 percent is pinyon pine (Pinus monophylla Torr.) -2- ,.,J· Figure 1.--Test fire plot before ignition. pickup trucks (right). Note Helicopter (left) and four CJ Tra iler LEGEND • Shielded thermocouple tower • Aspirated thermocouple tower • Non-shielded thermocouple tower I Ice bath (buried) s Soil prob e ~ 1£1 • eT- 2 tlil \. \l 1 1\!¥~ r _ Vacuum generator Wind direction I· 3 mph Figure 2 . --Plan l ayout of t es t -p l ot 760-2 showing the thermocoupl e tower l ocations. -3 - and 5 percent Utah juniper (Juniperus californica Carr. var. utahensis). The piles are 47 feet square and 7 to 10 feet high. They are spaced either 25 or 115 feet apart, depending on the type of plot. Care was taken to keep as much of the finer fuel intact as possible during construetion.1 The data reported here were recorded from fire in a 4-1 /2-acre plot containing 36 piles spaced 25 feet apart. Total fuel weight was 720 tons. The fuel loading for the entire plot was 170 tons per acre or 8. 0 pounds per square foot. The plot was constructed from July 1 to August 2, 1963. At that time fuel moisture was 86 percent on a dry weight basis. Average fuel moisture at the time of ignition, 8:30 a.m. May lq, 1964, was 20 percent. The heat of combustion of this fuel, as determined by standard oxygen bomb calorimetry, is 9, 500 B. T. U. per pound for needles and 8, 500 B. T. U. per pound for twigs, branches, and stems. Temperatures on this burn were measured with probes made from inconel-sheathed and magnesium -oxide insulated, 22 gauge (B. S.) chromel-alumel thermocouple wire, strapped to 1-1 I 2 -inch steel pipe towers. These probes corrected many of the prior problems encountered in supporting and insulating bare thermocouple wire within and above a field test fire. Plastic coated chromel-alumel wire led from the probe underground to the respective cold junctions. The pile probes extended to the edge of the fuel. The cold junctions were immersed in buried vacuum bottles filled with ice and distilled water. ) 'ot:rtil Copper wire was used as lead for the several hundred feet from each ice bath to instrument trailers. These lead-ins went through one of 3 sequence programers to balancing potentiometer recorder~. . The's~ were Varian Model No. G-llA, and had a !-second time constant. Each programer sampled 14 thermocouples in sequence and any given junction was sampled every 35 seconds for 2. 5 seconds. There were three types of thermocouple towers on this test btirn. Within the piles of fuel at 7 and 20 feet above the soil were bare chromelalumel thermocouples. In the aisles at the same heights were both shielded unaspirated and shielded a~pirated thermocouples to measure gas temperatures (fig. 2). One aspirated thermocouple was also located at 50 feet on the center tower. The aspirated system included thermocouple probes, .aspiration tubes, radiation shields, vacuum pumps, condensers, butane tanks and generators (figs 3, 4). The probes were fabricated from the previously described material and inserted just inside the aspiration tube. Galvanized steel thin wall conduit and compression coupling9 made up the vacuum manifolds. The radiation shield was galvanized sheet steel. A condenser and particle trap protected the pumps. 1 vJ D e t a i 1 s o f p 1 o t c o n s t r u c t i o n a r e g i v e n i n s t u d y· . p 1 a n N o . 2 5 2 1 E , on file at Pacific Southwest Forest and r:ange Experiment Station, Forest Service, U.S. Department of l.griculture, ~iverside, Calif. -4- Te lescoping Towe r ,--, ~~~L Fuel Compression coupling !teet L ·c---~~-~ To rec order ~~· __ Base di~t~ll~d l -i nch tube bur ied 1~ 1n che s water ~ -- Vacuum 1 I I L __ _j L __ ------J Figure 3.--Aspirated the rmocouple system showi ng the 50 foot tower. ~ Tower rP robe I _&;/~,....;;~~/ Shie Id r J ' ,.--.....r1--n>lc1hn--rii (p_:_ --=--=-~--t-L-f 3tB' l f Set sc rew\ ~~~~~ ~I ~ II ~~ I I 00 0000 1111 l -in . t hin wall 1= ~ tub ing }®C t t/ lf2·in. thin wall tubing V /114-i n. Q= t t hin wall t ubing (stainless stee l) 1;6 =t) . rCa lc iUm sulfate I . A1r gap Fi gu r e 4 . --Thermocouple shi e ld used on thi s t est fire. -5- Soil temperatures were obtained from probes at depths of 3 and 6 inches, between piles and under a pile. The junction was coated with plastic cement to insulate it froin ground. The shielded unaspirated thermocouples were protected from radiation by a multi -layer shield which consisted of three sections of tubing, each 4 inches long, secured one inside the other (figs. 3, 4). The outer tubes were 1- and 1/2 -inch galvanized steel. The smaller inner tube was 1/4 -inch stainless steel. A layer of calcium sulfate was placed between the 112 -inch and 1 I 4 -inch tube. A horizontal shield of galvanized sheet steel was placed 3 I 8 inch above the upper opening. · To ignite the fuel, ten 1-pound igniters of jellied diesel oil encased in plastic film were used in each pile. All igniters were fired simultaneously by means of electrical squibs and fuses. Ignition time (time zero) is the time when the fuel actually begins to burn (fig. 5). Results In all three of the instrumented piles, temperatures were recorded above 2650° F. However, chromel-alumel ·thermo.couples are not very·· reliable above 2300° and the materials melt at about 2600°F. Thus, recorded temperatures higher than 2600° , are of questionable accuracy. Temperature recorded at tower T-9 (see fig. 2) rose to 1140° F. at 20 feet ·and to 2204° F. at 7 feet in the first 6 minutes after ignition (fig. 6); the recorded temperature at 20 feet reached 2650° F. at 7-112 minutes. This tower began to fail at 13-3 I 4 minutes, and the heights of the thermocouples after this are not known. Recorded temperatures at T-10 and T-11 reached 2650° F. within 2-1 I 4 minutes after· ignition (figs. 7, 8); No valid data were recorded after this time from these two pile towers. Gas temperatures recorded with the aspirated system showed a similar quick jump from ambient to the initial maximum,. but there w.as a lag behind the values of the bare thermocouples. This lag can only be determined approximately because of the 3 5- second interval between readings, but appears to be 1 minute. The temperature at 20 feet was higher than that at 7 or 50 feet on tower T-4. The first peaks in gas temperatures (uncorrected) at 7, 20, and 50 feet were 1250° F., 2000° F., and 540° F., respectively (fig. 9). The initial peaks for tower T-3 at 7 and 20 feet were 1400° F. and 1100° F., respectively (fig. 10). The gas temperatures at T-3 and T-4 were quite similar at the second peak: about 1500° F. at 20 feet and 1300° F. at 7 feet. These two peaks are apparent in both gas and bare thermocouple data. Radiation measurements taken on the perimeter of the fire show also these two peaks at corresponding times. No data were recorded from T-1 and T-2 because of circuit failure. -6- Fi gur e 5.--Test fir e 3 minutes after firing igniters . -7- ~ 0 7-feet 18 C) C) --e ~ :::. Cb Q. e ~ I I 20-feet__, I I I. 300 600 Time after ignition (seconds) Figure 6.--Temperatures of bare thermocouple probes at T-9. 27~--~----~--~--~ 1 1 , Faii 1 Fail I I I I I ~ 181C) C) - I I - I I - 1 -7-feet - r- tJ 'Cb r- I - ~ 9r- I I I - ~ :::. ~ 'I'' - r-20-feet--1 ~ I I - I I '' ,___ '' 20-feet I 'I'' ''' ' I I I I I ' - I I 0 T-I 0 !!I'W'....,......,..,,1 - - -.- I ...1 -100 0 200 Time after ignition (seconds) 0 200 400 Time after ignition (seconds) Figure ?.--Temperatures of the bare thermocouple probes on T10 at 7 and 20 feet. Figure B.--Temperatures of the bare thermocouple probes on T11 at 7 and 20 feet. -8- 20 • ,.... --, ~15 I I,.......__ 0 a a --- II <1> ' \20-feet -..............'\ '- :::> ..._ ~10 <1> c.. ... . E: ~ ... \ -.,~7-feet \ .. .. ... ...... ·..... ·.. .... ... .. .~50-feet .... I • ~ 5 . 'l: I! ...........~-­ ....... . .... ... ------ t, ---......... ........ .... .... . 0~--------~--------L---------L---------L---------~--------~--~ -5 0 5 10 15 Time after ignition (minutes) 20 25 Fi gur e 9 . --Gas tempe r atu r e a t T-4, 7 , 20 , a nd 50 f ee t . 15 ~--------~--------~--------~---------,----------r---------~--~ "' \--- 20-feet ' ~ 0 a10 \ ' '"'""', \ \. a --- \. \. ~ \ :::> ..._ \. \ C) '- \,. <1> Q E: 5 ......... ............ ~ 0 ~--------~--------~--------~--------L---------L---------~--_J -5 0 5 10 15 Time after ignition (minutes) 20 Fi gu r e 10 . --Gas tempe r a t u r e on t owe r T- 3 a t 7 and 20 f ee t. -9 - 25 The error in gas temperature measurement by this aspirated system can be estimated from the following formula (Fishenden and Saunders 1939): Tg- Tc T _ T = N de. c w Where: T = temperature of the thermocouple c Tw = temperature of the shield wall T g =-· temperature of gas e d N = emissivity of chromel-alumel = diameter of wire in inches = relationship between gas. velocity~ diameter of wire and the mean of T and T , given c w in table by Fishenden and Saunders. The average of the gas velocity over the five aspirated couples was 56 feet per second as determined from a multivaned flowmeter and corrected for pipe friction as shown in Machinery's Handbook (Oberg and Franklin 1930). Assumed emissivity of chromel-alumel is 0. 5 based upon Fishenden and Saunders. The diameter of the wire is 0. 0201 inch and N is 35. The temperature of wall was not measured so an assumption must be made. At this gas velocity~ 200° F. difference between the wall and the junction seems reasonable. , If this assumption is used as a maximum the expected error is around + 4 percent. This means that the gas temperatures in Figures 8 and 9 are probably within t 4 percent of actual values. The position of the thermocouple and the high gas velocity greatly limit the effect of wall temperature on the junction: ...) The shielded unaspirated thermocouples~ all of which were towar.o the outside edges of the plot~ showed much lower temperatures than the aspirated thermocouples. An initial rapid rise is apparent at location T-5 and T-7, but does not show at T-6 and T-8 (figs. 11, 12, 1~, and 14). Readings at T-7 reached a maximum of about 1400° F. Readings at T -6 and T -8 never went above 500° F. Soil temperatures in the center of the plot (midway between the · four center piles) rose quite rapidly from ne.ar freezing to .120° F. at 3 -inch depth. The temperature at 6 -inch depth at this same location, rose from 32° F. to 60° F. in 45 minutes, then rose gradually to 75° F. at 180 minutes after ignition. It remained quite steady fqr ·some time after this (fig. 15). The temperature at 3 -inch depth between pil:es• · · decreased slowly after 30 minutes; 5 hours after ignition it was about . equal to the value at 6 inches. The soil temperatures under the pile of burning fuel began to rise 30 minutes after ignition and were still rising slightly after 5 hours and 40 minutes of recording. The temperature at 3 inches reached 17 5° F. ; at 5 inches, 105° F. The 30 minute lag behind temperatures between the -10- ._J ~15 t----7- feet 0 - 25 10 15 20 Time after ignition (minutes} 5 30 35 Figure 11.--Results from the shielded thermocouple probes on T-7. ~ I I I I I ---- , ____ --, _ ------ c g5 - I I - £20-feet Cb ""' .....::s e , / - 7-feetJ --------- ------ Q) Q. roo ~0-5 I I I 0 5 I I I 15 20 10 Time after ignition (minutes} I I I 25 30 35 Figure 12.--Results from the shielded thermocouple probes on tower T-8. ~ c S2 ......_ r 5 t- ~ .2 E Q) Q. E ~ ~5 ----T0 T T I I 0 I I I - 20-feet /7-feet , , .... .L,_. . . . . / ---- --r--------- - , , - -/ I 5 I I I 10 15 20 Time after ignition (minutes} I I I 25 30 35 Figure 13.--Results from the shielded thermocouple probes on tower T-6. 0 5 10 15 20 Time after ignition (minutes) 25 30 Figure 14.--Results from the shielded thermocouple probes on tower T-5. -11- 35 Figure 15.--Soil temperatur es at two dept hs, under a nd between pil es . piles is probably due to insulation of the soil surfac e from radiation and condu ctio n by non-burning fuel s uch a s large logs. Summary a nd Conclusions T emperatures of 265 0° F . were recorded by bar e t hermo couples in and above the flam e zone on a large field test f ir e . Eve n higher ternperatures are probabl e, judging from such evidence as the melting of steel and c h r omel w ir e and the light yellow - orange areas of flame appear ing in cer tain areas of the fire. This means that t h e 1500° F. measured by Bruce, Pong, and Fons (1959), th e 1800° F . predic t ed by Byr am (Davis 19 59 ), t h e 1500° F. reported by L i nd e nmuth and By ram (1948) , and 2000° F. predicted by Ve hrencamp (1956) are low for mass f ire situations . Th ese values re l ate to mu ch light er fuel lo adings t h an i n the present study. Our readings are more in agreement with t h e 2600° F. r ec orded by Broido and Mc Masters (1960). Ther efore, it may be necessary to use values of 2600° F. o r h igh er when trying to d esc ribe t he t hermody namics of large fires. Soil temperatures are in general agreement with past work although compar ison is diffi c ult because of differ ing te s t conditions. Two temperature p eaks were recorded by bar e and aspir ated the rmocouples in this fire . Two peaks have shown up on most previous test fires in thi s study and are a lso apparent from the r adiation data. The caus e is appar ently an initial fast rate of combustion of fine fuels foll ow ed by a period of les s intense burning until intensity of fire in t h e medium and coarser fuels reaches a maximum. Chrom el - alumel thermocoup l es ar e suitable for most temperature measurements on large fires . But to r ecord temp eratures abov e 2650° F., some oth er material suc h as platinum -platinum rhodium junctions , must be u s ed. An a spirat ed t h ermocouple s ystem proved pr acti c al on this large field -t e s t fire, but m e as urement of tube wall temp eratures would aid -1 2 - correction of gas temperatures for t he effects of the environment surrounding it. The unaspirated shielded th ermo coup les , produced much too great a lag time to be of value for gas temperature measurement, as Vehrencamp (1956) found in earlier work. E ar lier Forest Servi ce tests had shown that temper atur e records from bare th ermo coup les fluctuated so rapidly that comparison of hea t out put at different points in the fire was difficult (Countryman 1964) . B udget limitations prevented t he installation of heatingrate transducers, so t he shielded cou ples were used f or a rough measurement of heating at different locations. Apparently t he diffe rence in heat ing between the towers at T-5 and T-7 a nd those at T - 8 and T - 6 is due to inflow of ambient air. The two locations exhibiting higher hea ting wer e expos ed t o inflow fr om only o ne direction. Po ints T-8 and T-6 were exposed to inflow fr om two directions. In future tests the number of sample points will be increased to permit determination of hori zontal a nd vertical temperature pattern s. Heat flux will be measured with thin fo il heating-rate transducers . A self contained aspirated the rmocoup le is also be ing develop e d to simplify ins trumentation of large test fires. L iterature Cited B eadle, N. C . · W. 1940 . Soil t e mp eratures during forest fires and t h eir effec t on the survival of vegetation. J. E col. 28: 180- 19 2, illus. Bentley, J. R., and F enner, R . L . 19 58. Soil t emperatures during bur ning related to postfir e seed b eds on wood l and range. J. Forestry 56 : 7 37 - 7 40, ill us. Broido, A. , a nd McMasters, A. W. E ffect s of mass fires o n personnel in shelters. U.S. 196 0 . F orest Serv. Pacific SW. Forest & R ang e E xp. Sta., Berk e l ey, Calif. T ech . P ape r 50 . 89 pp., illus . Bruc e, H. D . , Pong, W. Y . , and Fo ns, W . L. Project fire model, 3rd progress report . U.S. Fores t 1959 . Serv. P acific SW. Forest & Range E xp. Sta., Berkeley, Calif. 23 pp., i llus . Countryman, C. M . 1964 . Mass fires and fir e behavior. U.S. Forest Serv. P acific SW. Forest & R ange Exp. Sta., B er keley, Calif. R es. Paper P SW -19. 53 pp., illus. Davis, K enn eth P . 1959. Forest fire: con trol a nd use. -13- 584 pp., illus. New York. . Davis, L. S., and Martin, R. E. :_ 1960. Time-temperature relationships of test head fires and backfires. U.S. Forest Serv. SE. Forest Exp. Sta., Asheville, N.C., Res. Note 148. 2 pp., illus. Fahnestock, George R., and Hare, Robert C. 1961. Temperatures of tree trunk surfaces and surrounding air in forest fires. Progress Report, U.S .. Forest Serv. Southern Forest Exp. Sta., New Orleans, La. 36 pp., illus. Fishenden, M. , and Saunders, 0. A. 1939. The errors in gas temperature measurement and their calculations. J. Inst. Fuel (London). 12(6): S5-S14, jllus. Heyward, Frank D. 1938. Soil temperatures during forest fires in the longleaf pine region. J. Forestry 36: 478-491, illus. Lindenmuth, A. W., Jr., and Byram, George M. 1948. Headfires are cooler near the ground than backfires. Control Notes 9(4): 8-9, illus. Fire Nelson, R. M., and Sims, I. H. 1934. A method for measuring experimental forest fire temperatures. J. Forestry 32: 488-490. Oberg, Erik, and Jones, Franklin D. 1930. Machinery's Handbook, Eighth edition, New York, The Industrial Press. 1464 pp. Vehrencamp, John E. 1956. An investigation of fire behavior in a natural atmospheric environment. Univ. Calif. Dep. Eng. Tech. Rep. 1. 85 pp., ill us. The A u t h o r - - - - - - - - - - - · - - - CHARLES W. PHILPOT is studying the chemical and physical ·characteristics of wildland fuels as part of the Station's research on the factors that influen~e forest fire behavior. He joined the staff in 1960. He received bachelor's and ma~~er'~ of scien~e degrees in forestry•from the Un1vers1ty of Cal1fornia in 1961 and 1962. -14- ...) ,.